iShowcase

UNDERGRADUATE CAPSTONE PROJECTS

The following InfoSci undergraduate capstone projects are eligible will be presented at the iShowcase and are eligible for awards.

Spring 2025 Undergraduate Capstone Projects

GAMES

Advanced Investments
Team members: Irene Leon, Hanwen Liang, Julien Nygaard, Conner Pessin, Keala Goodell
Advanced Investments is a strategy card-based game where players take on the role of a bank's AI assistant, using cards to provide financial advice to customers. The game aims to teach players financial literacy by applying financial decision-making skills to simulated real-life scenarios.

Dark Road to Mead
Team members: Doodle Biehle, Ben Bruso, Dillon Dye, Ethan Thomas and Jaden Dye
Dark Road to Mead is a survival horror game where the player maintains a car and avoids cryptic creatures while driving from the outskirts of Phoenix to the Hoover Dam.

Curtain Call
Team members: Vio Stavropoulos, Anh Thu Dang, Jonah Kirby, Mary Tran, Luis Ugarte
Curtain Call is a murder mystery narrative game with puzzle mechanics. An idol group, ACT: LOVE, finds themselves at a haunted house for a special episode on a variety show for up-and-coming artists. The night started fairly quiet until the group was given their first mission—and then… someone ended up dead. Trapped at the location with no way home, it is up to Juno to find a way to survive, solve the mystery and identify the killer before they strike again.

Upkeep
Team members: Egan Putman, Genesis Galilea Herrera, John Turpin-Cruz, Andreas Piyis and Jason Walls
A cozy medieval game where players clean, decorate and transform three distinct homes. Armed with an array of medieval-themed furniture, customization tools and a relaxing atmosphere, Upkeep offers players a satisfying experience where their creativity can run wild.

Warriors of the End
Team members: Santiago Crisantes, Julianna Dagenais, Yida Fang, Zachary Kosmerchock and Wenkang Lyu
You are a mercenary named Slythe, and you have been hired by a group of archeologists to guard them as they do research in an old temple that has been overturned by thieves. However, when you arrive at the temple an evil presence makes itself known. Play this JRPG inspired by the classics as you explore this temple, discover lost history, discover hidden secrets and discover what lies at the bottom of the ancient temple.

SOFTWARE

Bear Down Tracker
Team members: Claire Baker, Abby Cowden, Evya Sivarajan
Bear Down Tracker is a mobile app designed for the University of Arizona's community that can be used as a source for campus event tracking. It features a comprehensive events dashboard for events on campus tailored to the user’s interests. The app also offers customizable search filters to sort events by categories such as volunteer opportunities, tickets required, basic needs services and more. Its user-friendly interface makes it accessible and easy for students, faculty and staff to stay informed and involved with campus events.

Catch Flights Not Feelings
Team members: Ella Jacobson, Uyi-Osa Irowa, Ibrahim Rafiq, Rashid Abdi
Our project is a website that consists of advanced machine learning models to accurately predict flight delays. By leveraging flight data, weather conditions and external factors, we aim to create a comprehensive tool that helps passengers plan their travel more efficiently.

MockSphere
Team members: Sherali Ozodov, Khamdam Kadirov, Khondamir Rasulov and Sarvarbek Usmonov
MockSphere is a web-based platform that helps people practice for job interviews. It lets users do mock interviews through video calls with AI interviewers, live coding exercises and job-based questions. The platform also gives detailed feedback to help users improve. Recruiters can create their own interview flows and get reports on how candidates perform. Our goal is to make interview preparation easier, more realistic and more personalized.

OnQueue
Team members: Jackson Grove, Fuad Uddin
A mobile app that helps users reclaim their time by delegating tasks to AI agents. Users schedule tasks directly on their calendar, prompting AI agents to autonomously complete these tasks at the designated time, allowing users to focus on what matters most.

Wildcat Dining
Team members: Rachel Schmidt, Joseph Roth, Alina Suárez, Paloma Vasquez, Anthony Vo
The Wildcat Dining app is designed to help University of Arizona students navigate on-campus food options based on their class schedule and dietary preferences. With over 30 dining options across four dining districts, students can easily explore, filter and organize meals into a daily planner. Students can engage with their campus community by sharing and reading restaurant information and meal reviews, making it easier to plan meals and reduce stress on busy days.

RESEARCH

Galápagos Islands Testudinidae Macroevolution
Team members: Paige Cherry, Sara Cielaszyk, Hubert Kasprzcki and Eleanor Tuck
Did you know that some tortoises live for over 100 years? How do you think their long lifespan affects their evolution and adaptation? Our project looks at the macroevolution of tortoises as a part of the Testudinidae family. This isn’t just about tortoises—this is about how life itself evolves and diversifies over time, leading to the incredible variety of species we see today.

Healthcare Desierto
Team members: Sarah Ortiz, Chi Thieu, Danielle Cunes
Identifying how public transportation in Tucson affects healthcare access is essential to understanding where healthcare deserts are present in the city. By visualizing how patients access healthcare provider by public transportation such as Sun Link and bus, we are able to locate any underserved areas and bring awareness to both residents and the city.

Major League Baseball: A 2023 Rule Change Analysis
Team members: Alec Fernandez, Marissa Ronquillo and Alvaro Borbon
In 2023, Major League Baseball implemented several key rule changes aimed at improving the pace of play which has the goal of enhancing the fan experience and increasing both viewership and attendance. By analyzing in-person attendance trends before and after the implementation of these rules, we explore whether these changes have had a lasting impact and what the potential future for baseball could look like.

Psychology in eSports
Team members: Jaythan Baythavong, Kevin Li, Steven Aske, Tom Oswald, Rachel Stienstra
What we are testing in our project is whether or not winning early game rounds in competitive games affects the mental game of the players and whether they lead to winning in general. We have an R-shiny app that will go along with a research paper, showcasing our data and what has been found to be true or not true.

SonoraSafe
Team members: Buddy Buttram, Matthew Harper, Nicholas Kaplan, Max Stec
SonoraSafe is a research project focused on bicycle safety in Tucson. Our goal is to determine the most crash-prone intersections and high-risk types of traffic infrastructure for cyclists to avoid through quantitative analysis of bicycle crash data and qualitative analysis of intersections.

MASTER'S CAPSTONE PROJECTS

The following InfoSci Master of Science in Data Science and Master of Science in Information Science capstone projects will be presented at the iShowcase. Projects in the Student-Driven category are eligible for an award.

Spring 2025 Master's Capstone Projects

STUDENT-DRIVEN

AI-Driven Algorithmic Trading: Optimizing Market Strategies
Mentor: Tyson Swetnam
Team: Aman Singh, Chaithanya Konda, Hemant Kumar BK, Rohith Velan Singaravelu, Sanjeevteja Ponugumati
This project develops an AI-driven intraday trading system for stocks and ETFs in the AI and EV sectors, combining TimeGPT forecasts with real-time news sentiment analysis. Using Backtrader for backtesting and robust risk management, the system blends quantitative and qualitative signals to support smarter trading decisions.

An exploratory approach to AI-Powered Texture Synthesis for Interior Design
Mentor: Mathew Briggs
Team: Hamsini Achyutha Kumar, Meghashree N, Niharika Chavan Tatoji Rao, Shreya Rajendra Kadole
This project aims to create an AI-powered tool that transforms room images by applying realistic texture changes to furniture and surfaces. Leveraging object detection and generative models like Stable Diffusion and ControlNet, it enables material-based transformations while preserving lighting and spatial consistency. Produced within the AI and XR VIP. 

Data the Explorer
Mentors: Ash Black, Mathew Briggs
Team: Kashyap Sai Prasad Nadendla
Data the Explorer is an AI-powered analytics platform that transforms a user's data into intelligent, interactive visualizations using a multi-agent system. When a user uploads a dataset, agents automatically generate charts, analyze patterns, and review trends—creating contextual awareness of the data within the app. This context allows users to query their data conversationally, get tailored insights, and explore trends in real time. The platform offers a collaborative, intuitive experience—enabling anyone to uncover insights and make sense of their data. Produced within the AI and XR VIP. 

End to End MotoGP/F1 Live analytics Dashboard
Mentor: Tyson Swetnam
Team: Naveen Saragadam, Rohith Gadichanda, Likhith Ramesh
This project focuses on building an end-to-end analytics pipeline that ingests live data from public APIs, processes it, and presents insights via an interactive dashboard. Potential data sources include Formula 1 telemetry and domestic flight data, with an emphasis on automation, cloud tools, and data storytelling.

FashionFlux
Mentor: Maggie Keef
Team: Himanshu Lal, Ajeet Singh, Aastha Prasad, Bhaskar B. Kashyap, Viren Sasalu
This project aims to develop an AI-powered virtual try-on platform using Stable Diffusion and ControlNet to generate realistic garments over user images in real time. Designed for e-commerce and social media, it enables seamless online shopping and empowers influencers to create and showcase outfits without physical clothing.

LiveBreak
Mentor: Michael McKisson
Team: Raja Sekhar Malireddy, Anand Ramaswamy Jayshree, Vishnu Rendla, Reshma Sai Yarlagadda, Sreeharsha Nalluri
LiveBreak is a hyper local news delivery platform designed to provide users with personalized credible, and engaging news experiences. The platform integrates cutting-edge artificial intelligence techniques to enhance content creation, verify news authenticity, optimize user engagement, and deliver news via text and speech.

Multi-Purpose AI Agent Framework
Mentor: Tyson Swetnam
Team: Prudhvi Kandregula, Vamsi Krishna Chagarlamudi
This project develops a Multi-Purpose Agentic Flow Framework that allows users to build customizable workflows by connecting intelligent agents for tasks like document processing, API integration, and database querying. With a modern web interface and containerized deployment, the system supports scalable, reproducible, and efficient pipelines for both structured and unstructured data.

Optimizing Federated Learning in LLMs
Mentor: Jyothikrishna Dass
Team: Hariharan Ramesh
Our project, FLoRIST (Federated LoRA with Singular Value Thresholding), introduces an advanced method for federated fine-tuning of Large Language Models (LLMs). It optimizes communication efficiency and scalability while maintaining mathematical accuracy in model aggregation.

Quantifying Cognitive Bias in Large Language Models Through Decision-Making
Mentors: Tyler Millhouse, Cristian Roman-Palacios
Team: Anudeep Appikatla, Venkata Satya Murali Krishna Chittlu
Our project aims to develop a quantitative mathematical framework to assess and measure bias in LLM decision-making. By creating structured prompts and leveraging mathematical techniques, we seek to objectively analyze how different LLMs make decisions, allowing for transparent and comparable evaluations across models.

Self-hosted AI Applications Using Domain-Specific Custom Context to Accelerate Bioinformatic Research in Genomics
Mentors: Egoitz Laparra, Bernardo Lemos
Team: Aslam Sheik Dawood, Jose F Oviedo, Pranshu Singh Rawat, Vamsi Vadala, Visalakshi Iyer
This project builds a domain-specific Retrieval-Augmented Generation (RAG) chatbot to assist researchers in single-cell genomics. By combining open and closed-source NLP tools, the system enables semantic search, metadata queries, and detailed document analysis to streamline research workflows through an intuitive web-based assistant.

Self-rewarding deep reinforcement learning for trading strategy optimization
Mentor: Greg Chism
Team: Sohan Shankar Arasavilli, Tanmay Naik, Vishnu Rendla
This project develops a Self-Rewarding Deep Reinforcement Learning (SRDRL) trading strategy that adapts to dynamic financial environments by learning its own reward signals. By combining supervised reward estimation with expert financial metrics, the model aims to optimize decision-making through a novel Self-Rewarding Double Deep Q-Network (SRDDQN).

Summiva: An Enterprise-Scale NLP System
Mentor: Liangming Pan
Team: Disha Motwani, Saikumar Bollam, Sivarajan Jaganathan
This project focuses on Summiva, an enterprise-scale NLP platform that summarizes, tags, clusters, and retrieves unstructured text data using advanced transformer models and clustering algorithms. Designed for high-volume, ethical content processing, it combines Elasticsearch and FAISS to deliver fast, structured, and searchable insights.

FACULTY ONGOING PROJECTS

A statistics game in R & Unity
Mentor: Meaghan Wetherell
Team: Jonathan McCoy, Varun Jayaram
This project aims to recreate a hands-on statistics game—building and destroying histograms—using Unity and R to develop an interactive online version that helps students explore distribution structures in a playful, visual way.

Advancing LEO’s Phenocamera data pipeline and exploring vegetation dynamics of Biosphere 2 Landscape Evolution Observatory
Mentor: Wei-Ren Ng
Team: Lokeshwar Reddy Gowkanapalli
University of Arizona Biosphere 2 Landscape Evolution Observatory (LEO) research project is looking to expand its remote sensing capability as preparation to transition to the ecology phase of the experiment. We continue our efforts to fine tune the work started last semester to process geo spatial datasets/images and tools to help our efforts to track the change of its phenology.

AI for Predicting Healthcare Accessibility Gaps: A Multimodal Approach Using Public Data
Mentor: Greg Chism
Team: Akash Satpathy, Ansh Dev, Himanshu Manoj Nimbarte, Jay Patil
Develop a multimodal AI system that combines geospatial data, socioeconomic indicators, public health statistics, and transportation networks to predict and visualize healthcare accessibility gaps across underserved regions.

AI-Powered Polypharmacy Risk Predictor:
Mentor: Greg Chism
Team: Rishab Khatokar, Shashank Satish Kohade, Navyasree Madhu, Tushar Shrivastava, Harsh Wasnik
Our product, the AI-Powered Polypharmacy Risk Predictor, is a web-based decision-support tool designed to identify and predict potential adverse drug-drug interactions (DDIs) in polypharmacy scenarios. It integrates a machine learning model with an intuitive user interface to assist healthcare providers in evaluating patient medication regimens.

Analyzing joke comment behavior on PBS Eons YouTube videos
Mentor: Meaghan Wetherell
Team: Anubhav Mathur, Siddheshwar Singh Negi, Sumel Rattan
Analyzing YouTube comments to determine if there are lots of repeated/similar jokes on videos that went viral and whether leaving an obvious joke out of the episode would potentially make commenters more likely to post. 

Analyzing survey data assessing the impacts of audio sketches on belief and openness to civil discourse
Mentor: Diana Daly
Team: Kendall Beaver
This project involves analyzing a pilot survey on the impact of AI-generated audio and visual ads, with potential expansion to a larger study. Students will support data collection and analysis, but must have a strong grasp of the survey’s design and objectives to ensure meaningful results.

Census Data Analysis and Predictive Modeling
Mentor: Sean Kramer-Lazar
Team: Morgan Godley
This project examines prior census data gathering by the university regarding course enrollments. We looked at trends in course offerings, numbers of sections, modalities, and enrollments to identify trends, analyze emergent patterns, and create predictive models to help understand future schedule changes and modifications.

College of Information Science Graduate Admissions Trends and Data Visualization
Mentor: Tavia Szostek
Team: Ajay Sreekumar, Anushree Biswas, Naitik Shah, Panneer Selvam Mani Sekaran
This capstone project involves analyzing graduate admissions and matriculation data for the College of Information Science and producing data visualizations to highlight current trends. Students will support data processing, analysis, and presentation to inform future admissions strategies.

Data Analytics for Health and Safety
Mentor: Leonard D. Brown
Team: Yanyan Dong
This program uses NLP and gen AI to process reporting corpus in mine operators' safety management systems. The project seeks to identify leading indicators of health and safety, predict potential for injury, and suggest mitigation techniques for new job planning on the worksite.

Deceptive Dialogues: Exploring Gothic-Themed Deception in Large Language Models
Mentor: Greg Chism 
Team: Aditya Bandimatt, Durga Prasanth Gubbala, Giridhar Kotha, Nithin George Mathew, Siva Bhargav Ravula
This project investigates how large language models (LLMs) can simulate deception through dialogue, drawing inspiration from gothic characters like Dracula. Students will design and analyze chatbots with deceptive personas, examining how language, tone, and narrative structure contribute to trust, misdirection, and manipulation. The project blends computational linguistics, literary tropes, and AI ethics to explore the darker edges of human-AI interaction.

Deep Learning Pose Estimation
Mentors: Carlos Lizarraga-Celaya, Nirav Merchant
Team: Adrian Girone, Nick Ferrante
This project uses DeepLabCut, an open-source pose estimation tool, to track animal behavior in dynamic environments without invasive markers. Students will train neural networks to analyze behavioral videos, supporting research across fields like neuroscience, ethology, and biomechanics.

Enhancing Logical Reasoning in Large Language Models
Mentor: Liangming Pan
Team: Usama Ahmed
This research project builds on prior work to enhance the logical reasoning abilities of Large Language Models (LLMs) using a synthetically generated logic corpus. Through controlled experiments on training parameters and dataset design, the team aims to improve LLMs’ consistency and inference capabilities for applications like theorem proving and AI-driven decision support.

FastSpectralNet
Mentor: Eung-Joo Lee
Team: Jayant Biradar, Hitarth Bharad
Our research project introduces FastSpectralNet, a novel model designed for the classification of Hyperspectral Images (HSIs), which are used in remote sensing applications such as agriculture, environmental monitoring, and urban planning. HSIs contain rich spectral and spatial information, but traditional deep neural networks like CNNs and RNNs struggle to efficiently handle their high-dimensional nature.

Gamification to Improve Health and Safety Training 
Mentor: Leonard D Brown 
Team: Sankalp Sethi
This project seeks to create an AI-enabled serious game to teach hazards recognition, mitigation, and operating procedure best practices in health and safety for the mining industry. We are developing a multi-user virtual world with genAI-enabled non-player characters in Unreal Engine 5.

Generative AI for unique cultural heritage 
Mentor: Xiao Hu 
Team: Aditya Jolly, Maria Nikitha Suresh
This project tasks students with fine-tuning or building a generative AI model to create images representing a specific cultural heritage of their choice. Using techniques like Stable Diffusion, the goal is to address the underrepresentation of certain cultures in GenAI outputs by developing a tailored model and documenting the process.

Humility in scientific inquiry 
Mentor: Sarah Bratt 
Team: Akash Satpathy
This project involves analyzing scientific articles to identify humble phrases using a predefined codebook. Students will assist with text annotation, processing, and further analysis to support and extend an existing well-performing model.

Improving Navigation and Accessibility for Visually Impaired People at the University of Arizona
Mentors: Carlos Lizarraga-Celaya, Nirav Merchant
Team: Rohit Surya AVB, Abhishek Deore, Sahana Santhosh, Hema Priya Prakash
This project enhances the University of Arizona campus map by generating textual summaries of buildings and surroundings to support users with visual impairments. Summaries include addresses, landmarks, and accessibility features to improve spatial understanding and accessibility.

Process analytics of human-AI collaboration 
Mentor: Xiao Hu 
Team: Tallapaneni Venkateshwara Chowdary
This project will apply various process mining techniques on an anonymized dataset of human-AI collaboration, to reveal and compare interaction patterns. Deliverables will include code and a concise report with visualizations.

PubMed Agentic Retrieval Augmented Generation
Mentor: Enrique Noriega 
Team: Abhay Kumara Sri Krishna Nandiraju, Abhishek Kumar, Dhawal Gajwe, Syed Junaid Hussain
This project guides students in building an Agentic Retrieval-Augmented Generation (RAG) system using PubMed Central papers. Students will develop a full pipeline—from data ingestion and vector indexing to multi-hop agent-based reasoning—culminating in a working prototype, documented GitHub codebase, and a blog post summarizing the project.

Quantum for Good
Mentor: Sarah Young 
Team: John Kang
This project explores the rhetoric of "quantum for good" in preparation for a spring workshop with the Center for Quantum Networks. The focus is primarily non-technical, examining how language shapes public understanding and ethical framing of quantum technologies.

Scientific Collaboration Networks on Software
Mentor: Sarah Bratt 
Team: Swetha Kolloju
This project uses Python, R, and GROBID to extract full-text from scientific PDFs and identify software mentions, focusing on phylogenetic tools. Students will then conduct collaboration network analysis to explore collaboration patterns among scientists who use these tools.

UI research to support explainable AI in scientific data repositories
Mentor: Natalie Raia
Team: Christian Ortmann     
This project investigates what information scientists need to trust AI-curated geology datasets. Students will design interface wireframes for data repositories, conduct user testing, and explore how varying levels of AI method transparency impact trust and usability.

Virtual Harlem
Mentor: Bryan Carter
Team: Christian Ortmann
The Virtual Harlem Project is a collaborative learning network whose purpose is to study the Harlem Renaissance, an important period in African American literary history, through the construction of a virtual reality scenario that represents Harlem, New York, as it existed between the 1920-30s.

VR of Historical Places - Shared Churches after the 30-years war
Mentor: P. Bryan Heidorn
Team: Sree Krishnaprasad
This project invites students with experience in VR and simulation tools like Unity to help develop immersive experiences showcasing simultaneum mixtum churches—shared Catholic and Protestant spaces from the 16th century onward. Students will explore ways to present 360° imagery and metadata in both VR and web environments, enabling users to navigate spaces and access contextual information through interactive elements.

VERTICALLY INTEGRATED PROJECTS (VIP)

AI For Medical Interviewing
Mentor: Winslow Burleson
Team: Jasdeep Singh Jhajj, Shashwat Singh
This hands-on project explores how AI can enhance patient-provider interactions by personalizing care, improving feedback, and supporting provider reflection. Based at ASTEC, the team uses design thinking and draws from adaptive teaching, machine learning, and human-computer interaction to drive innovation in digital healthcare.

AI-Driven Healthcare Applications
Mentor: Eung-Joo Lee
Team: Joel Jojo, Avikal Singh, Libin N. George, Solman Raju Sarva
The AI-Driven Healthcare Applications team focuses on developing AI technology that can perform a range of basic tasks, from detection and recognition of medical conditions to more intricate challenges such as segmentation. The team specializes in developing deep learning systems specifically tailored for embedded computer vision and various healthcare applications.

Biosphere 2 Rainforest Resilience: Capturing and Communicating Ecosystem Responses to Climate Change
Mentor: Joost Van Haren
Team: Sai Laasya Gorantla, Monica Tejaswi Kommareddy
This project invites students to conduct hands-on climate change research in the Biosphere 2 rainforest, a uniquely controlled environment experiencing extreme canopy temperatures. Through seminars, internships, and fieldwork, students will explore tropical forest responses to heat and drought, examining carbon and water cycling, plant adaptation strategies, and greenhouse gas emissions using long-term data and active measurements.

Multimodal Emotion Recognition System using Florence-2
Mentor: Eung-Joo Lee
Team: Anjani Sowmya Bollapragada
This project aims to develop a Multimodal Emotion Recognition system using Florence-2 to classify emotions in conversational videos. By combining visual data and transcribed text from OpenAI's Whisper, the model provides deeper insights into affective states like joy, sadness, anger, and surprise.

Pedestrian Intention Estimation and Collision Risk Modeling
Mentor: Eung-Joo Lee
Team: Sai Manoj Chatrathi
This project focuses on the development of a pedestrian intention classification and collision risk estimation model using the PIE (Pedestrian Intention Estimation) dataset. The goal is to build a system that can first determine whether a pedestrian is moving or stationary and then evaluate the potential risk of a collision based on pedestrian trajectory, position relative to the ego-vehicle, and vehicle motion characteristics.

ADDITIONAL INFORMATION

For general information or assistance, please contact Jana Phillips, events coordinator, at jana42@arizona.edu or 520-621-8288.

For corporate queries, or to sponsor the iShowcase, please contact Michael McKisson, associate professor of practice and director of undergraduate studies, at mckisson@arizona.edu or 520-621-7556.